Objective: To examine the prospective associations of baseline abdominal obesity and TV-viewing time with five-year reductions in leisure-time physical activity level. Methods: We used data from the Australian Diabetes, Obesity and Lifestyle Study (AusDiab), a nationally representative population-based cohort study with measures collected in 1999-2000 and 2004-2005. Abdominal obesity was determined by waist circumference and TV-viewing time and physical activity level were assessed using established interviewer-administered questionnaires. Results: Among 2,191 men and 2,650 women aged ≥ 25. years, odds ratios (ORs) of 5-year reductions from sufficient to insufficient or no physical activity, and from insufficient to no physical activity were estimated with logistic regression. We adjusted for sociodemographic characteristics. The odds of reducing physical levels from baseline to the follow-up survey for obese men (34.2%) and women (38.1%), respectively were 1.40 (1.10-1.79) and 1.44 (1.16-1.80), compared to those with a normal waist circumference. Women, but not men, with higher levels of TV-viewing time had higher odds of reducing physical activity levels (8.6%; OR 1.46; 1.01-2.11), independent of abdominal obesity. Conclusions: These findings suggest that abdominal obesity is associated with prospective reductions in physical activity level, and that high levels of TV-viewing time might have an additional adverse influence for women. © 2011 Elsevier Inc.